Haplotype block structure and its applications to association studies: power and study designs

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• The human genome exhibits haplotype block structures with limited diversity.
• Tagging single-nucleotide polymorphisms (SNPs) can represent common haplotypes within these blocks.

Purpose of the Study:

• To quantitatively assess power loss in genetic association studies using tag SNPs.
• To compare the efficiency of tag SNPs versus all SNPs or randomly selected SNPs.

Main Methods:

• Simulation studies generated case-control and case-parental data.
• A dynamic programming algorithm identified haplotype blocks and tag SNPs.
• Statistical power was evaluated using all SNPs, tag SNPs, and random SNPs.

Main Results:

• Tag SNPs reduced genotyping by up to 86% with only a 4% average power loss.
• Compared to random SNPs, tag SNPs showed significantly less power reduction (4% vs. 12% at 25% SNP usage).
• Haplotype-based analysis demonstrated greater power than marker-by-marker analysis.

Conclusions:

• Tag SNPs offer a significant reduction in genotyping effort without substantial loss of statistical power.
• Tag SNP selection is a more efficient strategy than using all SNPs or randomly selected SNPs for association studies.
• Haplotype analysis is a powerful approach for genetic association studies.